Causal inference is a branch of quantitative research that attempts to rigorously estimate causal effects between phenomena of interest. For instance, we might ask if running regularly increases lifespan.
Broadly, the steps in a given causal analysis might be:
- Specify a causal question;
- Draw our assumptions about the relationships between the variables of interest using a causal diagram (a DAG);
- Collect data (if we don’t already have it);
- Model our assumptions (e.g. use propensity score matching);
- Diagnose our models (e.g. diagnose the effectiveness of our propensity score matching procedure);
- Estimate the causal effect;
- Conduct sensitivity analyses on the effect estimate